Molecular Cancer BioMed Central Review Open Access In silico gene expression analysis – an overview David Murray1, Peter Doran1, Padraic MacMathuna2 and Alan C Moss*3 Address: 1General Clinical Research Unit, UCD School of Medicine and Medical Sciences, Mater Misericordiae University Hospital, Dublin 7, Ireland, 2Gastrointestinal Unit, Mater Misericordiae University Hospital, Dublin 7, Ireland and 3Division of Gastroenterology, Beth Israel Deaconess Medical Center, 330 Brookline Ave, Boston, MA 02215, USA Email: David Murray - [email protected]; Peter Doran - [email protected]; Padraic MacMathuna - [email protected]; Alan C Moss* - [email protected] * Corresponding author Published: 7 August 2007 Received: 27 June 2007 Accepted: 7 August 2007 Molecular Cancer 2007, 6:50 doi:10.1186/1476-4598-6-50 This article is available from: http://www.molecular-cancer.com/content/6/1/50 © 2007 Murray et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Efforts aimed at deciphering the molecular basis of complex disease are underpinned by the availability of high throughput strategies for the identification of biomolecules that drive the disease process. The completion of the human genome-sequencing project, coupled to major technological developments, has afforded investigators myriad opportunities for multidimensional analysis of biological systems. Nowhere has this research explosion been more evident than in the field of transcriptomics. Affordable access and availability to the technology that supports such investigations has led to a significant increase in the amount of data generated. As most biological distinctions are now observed at a genomic level, a large amount of expression information is now openly available via public databases. Furthermore, numerous computational based methods have been developed to harness the power of these data. In this review we provide a brief overview of in silico methodologies for the analysis of differential gene expression such as Serial Analysis of Gene Expression and Digital Differential Display. The performance of these strategies, at both an operational and result/output level is assessed and compared. The key considerations that must be made when completing an in silico expression analysis are also presented as a roadmap to facilitate biologists. Furthermore, to highlight the importance of these in silico methodologies in contemporary biomedical research, examples of current studies using these approaches are discussed. The overriding goal of this review is to present the scientific community with a critical overview of these strategies, so that they can be effectively added to the tool box of biomedical researchers focused on identifying the molecular mechanisms of disease. 1. Background ing point in mechanisms of disease research is deciding Investigations aimed at deciphering the molecular events how to identify these disease-associated biomolecules. that underpin the initiation and progression of disease are primarily targeted towards the profiling of biomolecules, Historically such investigations focused on the characteri- whose aberrant expression, contributes to alterations in sation of single molecules and studying their role in dis- cellular function and ultimately lead to disease. By focus- ease. The inherent weakness of such focused disease ing on the mechanisms of disease, biomedical researchers research strategies lies in the fact that complex diseases are aim to identify critical molecular events that can be tar- usually polygenic and single molecule studies will not geted with novel therapeutic strategies. Thus, a key start- provide insights into the orchestrated response of a cell as Page 1 of 10 (page number not for citation purposes) Molecular Cancer 2007, 6:50 http://www.molecular-cancer.com/content/6/1/50 it evolves within a diseased tissue. Thus, it is accepted that type and also to as compare the expression of one gene in an overall view of the biomolecular composition of dis- various tissues or disease types. eased tissue provides extraordinary opportunities to observe the global molecular response to disease. By visu- Thus, biomedical researchers are equipped with both the alising the entire response, researchers begin to under- map of the genome and an understanding of how gene stand the complex inter-relationships between expression events contribute to health and disease. How- biomolecules that contribute to changes in cell pheno- ever, to truly capitalise on this wealth of information, type, and ultimately disease. A significant hurdle for bio- novel tools are required to permit identification of what medical researchers to overcome in the past has been how genes are activated and suppressed in disease. Techniques to access and analyse molecular information at such a capable of quantifying gene expression enable the devel- detailed level. The answer to this has been the develop- opment of our understanding of the distribution and reg- ment of novel experimental and analytical methodologies ulation of gene products in normal and abnormal cell that have, in many ways, redefined the biologists' toolkit. types. These include a variety of microarray and Serial Analysis of Gene Expression (SAGE) techniques, all of A major enabling factor in molecular analysis of disease which have the ability to quickly and efficiently survey has been the recent completion of the human genome genome-wide transcript expression. The development of project. This landmark project has detailed and defined microarrays has improved our ability to simultaneously our genetic make-up provides all the information needed study the expression of many genes in a particular tissue. to understand both health and disease. Although greeted However there are also opportunities to exploit computa- with much fanfare the completion of the genome- tional methodologies that profile expression of all genes, sequencing project is best seen as a new beginning for bio- not just known genes on chips, in a quantitative and medical research, as the sequence merely lists our genetic straightforward way. The availability of vast amounts of composition and does not interpret the relevance of the sequence data, coupled to advances in computational information in health and disease. However the availabil- biology provides an ideal framework for in silico gene ity of this data coupled with ongoing sequence determina- expression analysis. The last two decades have seen tre- tion initiatives has provided a huge repository of sequence mendous advances in computational approaches to data for use in assembly projects and also for enabling understanding the molecular basis of disease, advances continued developments in human transcriptomics, thus that have heralded a new era in biomedical research. The facilitating investigations of biological and disease mech- exponential growth of biologically relevant datasets has anisms to be carried out on a genome wide scale. transformed the biological and biomedical research enter- prise from a very data light to an information-heavy pur- All biological events in the cell are governed primarily by suit. This growth in available information has been changes in the expression of key genes. The ability of a cell matched by advances in our ability to understand and to switch on and off gene expression drives all biological mine this new information. Biologists now routinely ana- function and activity. Gene transcription is crucial in nor- lyse huge microarray datasets, recreate biological net- mal events such as cell division, proliferation, differentia- works, identifying protein folding patterns and model tion and cell death. Conversely, gene transcription is a whole cell activity using computational strategies. All facilitator of the pathogenomic events that drive the these advances are driven by computational strategies that development and progression of disease, as well as gov- match the availability of data, with the clear goal of iden- erning response to therapy. Much interest is therefore tifying biologically relevant patterns in data. Indeed these focused on the delineation of gene expression profiles to technologies have been used to investigate the molecular identify those key genes and gene clusters whose expres- events underpinning various malignancies, including sion is altered in disease states. Research into the mecha- breast, colon, lung, ovarian, pancreatic and prostate can- nism of diseases is underpinned by identifying these gene cers [1]. In this review a number of these strategies and alteration patterns. By comparing gene expression profiles their important, emerging roles in disease research are dis- under different conditions, individual genes or groups of cussed. genes can be identified that play a key role in particular signalling cascades or particular cellular process or in dis- 2. The assembly and organisation of in silico gene ease aetiology. Expression profiling is also important for expression data understanding gene functions and identifying therapeutic The growth in the number of EST mining projects is due targets. Gene expression profiling is also crucial to identi- mostly to the public availability of transcribed sequences. fying diagnostic, prognostic and predictive markers of dis-